News:

Precision Simulator update 10.173 (24 February 2024) is now available.
Navburo update 13 (23 November 2022) is now available.
NG FMC and More is released.

Main Menu

Sad: Atlas 767F down near Houston

Started by Jeroen Hoppenbrouwers, Sun, 24 Feb 2019 12:25

IefCooreman

TOGA will not work at that altitude, right?

Nightshifts have crazy effects one's brain... Never understood the flighttime regulations in the USA for freighter pilots. If this is true, this is a solid case to start reviewing this.

emerydc8

TO/GA arms with glide slope capture or flaps out of up.

UPS and its law firm, Jones Day, used the FAA Assistant Chief Counsel, Rebecca MacPherson, to sign off on the cargo carveout just before she left the FAA to take up a job as "Of Counsel" with Jones Day. When I called her out on it, she blamed the OMB's economic analysis. It's a revolving door between government and industry.

http://www.lawyerdb.org/Lawyer/Rebecca-MacPherson/

https://www.faa.gov/regulations_policies/rulemaking/recently_published/media/2120-AJ58-FinalRule.pdf

QuoteAll-cargo carriers may choose to comply with the new part 117 but are not required to do so.  Since the carrier would decide voluntarily to comply with the new requirements, those costs are not attributed to the costs of this rule. 

Actually, I truly hope fatigue is a causal factor in the Atlas crash. It wouldn't surprise me at all. I fell asleep at the controls last week due to a ridiculous (but legal) schedule.

Jeroen Hoppenbrouwers

Yes my mind has trouble understanding it now. Accidental full thrust and accordingly a pitch-up -- ok. But then a startled pilot would PULL the yoke to keep the airspeed in check?!

Or would everybody on the deck think they were  already climbing steeply, due to somatogravic illusion (accelleration feels like climbing) and having missed the engine oops due to for example noise of, say, thunder or rattling kit due to turbulence, no outside horizon, etc.?

http://aviationknowledge.wikidot.com/aviation:somatogravic-illusion

Need more info. It will come.


Hoppie

skelsey

Quote from: Jeroen Hoppenbrouwers on Sat, 16 Mar 2019 11:20
Yes my mind has trouble understanding it now. Accidental full thrust and accordingly a pitch-up -- ok. But then a startled pilot would PULL the yoke to keep the airspeed in check?!

Or would everybody on the deck think they were  already climbing steeply, due to somatogravic illusion

That was my initial thought and what is implied (at least to my reading) by the WSJ report (with phrases like "the seemingly disoriented crew"). Although it does seem strange given that the videos etc seem to suggest that it was largely day VMC at the time -- normally one would expect such sensory illusions to be more of an issue with no visible horizon.

Hardy Heinlin

Interestingly, the full-flight sims do just that: Lateral acceleration is simulated by pitching the platform nose-up.

Jeroen Hoppenbrouwers

767 drivers. Assume flight test. You fly 230 knots level at 6000 ft and intentionally firewall the throttles, count to five, and then agressively pitch down until 50 degrees nose down. How long would this take? Where are you at that moment? Speed and altitude?

Hoppie

IefCooreman

If they were looking outside when they pitched up and did not refer to the instruments, the pitch down at night can be killer. There are "visual" cases where you might not really be able to see when you "cross" the horizon or clouds might give you a completely wrong visual impression. You will not be able to tell when to stop pitch down. The pitch down value really makes me feel they had the impression they weren't going down yet. You really have to push long and hard to get that low pitch.

If you then look at the EADI, there will be a few moments where you are completely lost. That's when the famous IFR phrase "trust your instruments" should come into play. People tend to trust what they "believe" in those cases. This wouldn't be the first time somebody freezes up on this situation.

Hardy Heinlin

#67
Weren't there daylight VFR conditions?


Quote from: Jeroen Hoppenbrouwers on Sat, 16 Mar 2019 19:33
767 drivers. Assume flight test. You fly 230 knots level at 6000 ft and intentionally firewall the throttles, count to five, and then agressively pitch down until 50 degrees nose down. How long would this take? Where are you at that moment? Speed and altitude?

Will you find anyone who has tested it in the sim?

I guess from 6500 ft on it would take less than 30 seconds at a sinkrate way higher than 10000 fpm, at over 300 KIAS.

Can this be recovered at 4000 ft?


IefCooreman

For some odd reason my mind thought they crashed just past midnight... (which was why I referenced to flighttime limitations at night)

Will

In IMC, or at night with no horizon, I could imagine a pitch-over to -49 degrees in an inexperienced crew. But in visual conditions? The whole windshield would have been full of farms, lakes, roads and cars.
Will /Chicago /USA

emerydc8

I just happened to read this on pprune. I'm not making any assertion about it's veracity or accuracy, but it raises a few interesting questions. One is whether TO/GA will activate if pressed first and then the flaps are moved out of up a split second later. The other is whether the break-out box for the elevator will command a split elevator or if there is a "prevailing" side that would cause both elevators to work in unison. I'm thinking the elevators are bussed together with a torque tube. Later in the post someone also mentions that the speed brakes were out when this happened and the autopilot was engaged.

https://www.pprune.org/rumours-news/...l#post10425322

Quote"The initial bobble is from turbulence at 6200'. When the FO called for flaps 1, the captain accidentally hit the toga button. Toga didn't engage until after flaps were set to 1, which then brought engine power to full, and started the initial pitch of 10 degrees nose up. The FO was startled, and shoved the nose forward... The CVR is startling, and baffling. The CA was pulling so hard against the FO that he sheared the pins on the stick and at that point had no control. They were IMC at the time. When they broke out into VMC, the FO said oh schit and started to pull. That was the round out you see. I won't get into anything more until everything comes out. The records, the CVR, and what happened in the flight deck is truly shocking. They hit a negative 4 G dive initialy on the FOs push. All you hear is stuff hitting the ceiling and at one point a loud thud. They think the thud may have been the JS hitting the ceiling and maybe not wearing the shoulder harness. Like I said, I won't get into anything more about the background of how it all happened. This is the accident in a nutshell. The facts that will come out are shocking."

John H Watson

QuoteThe other is whether the break-out box for the elevator will command a split elevator or if there is a "prevailing" side that would cause both elevators to work in unison. I'm thinking the elevators are bussed together with a torque tube.

There is a torque tube in the tail, but there is an override mechanism on this torque tube which allows independent left/right side movement up to a 20° split.  The column force required to override is 25 lb at breakout, increasing to 41 lb at 20° of elevator movement.

QuoteThe CA was pulling so hard against the FO that he sheared the pins on the stick and at that point had no control.

My notes don't say what happens if the crew are opposing each other. It only says that 25lbs breakout force between the pilots (at the control column end) will decouple the independent cable system (via the "control columns override system" on the torque tube between the columns). There is no mention of shear pins in any of my manuals.

QuoteToga didn't engage until after flaps were set to 1

Bizarre.

For info:

With no air load, the elevator can move a maximum of 28.5° up and 20.5° down with a full forward and aft movement of the control column. The maximum movement of the elevator in autopilot mode depends upon the number of flight control computers engaged. In single autopilot, the maximum authority is 8.3° up and down. In multiple autopilot, the maximum authority is 28° up and 20° down.

emerydc8

QuoteThe maximum movement of the elevator in autopilot mode depends upon the number of flight control computers engaged. In single autopilot, the maximum authority is 8.3° up and down. In multiple autopilot, the maximum authority is 28° up and 20° down.

That's the max the servos will move it, right? If the A/P was engaged and a pilot manually pushed forward as hard as he could, would it still go to the normal limit (20.5)?

John H Watson

QuoteIf the A/P was engaged and a pilot manually pushed forward as hard as he could, would it still go to the normal limit (20.5)?

Correct. There is a "30lb force per channel" mentioned for servo "camout". I'm not 100% sure if this corresponds to pilot input. 90lbs force to override triple engage seems a little excessive.

John H Watson

Just noticed that the cutout switches which inhibit trimming when the columns are moved in the opposite direction to the column trimming on the 767 are activated at 2.2~2.7 degrees (for nose up trimming) and 2.0 ~ 2.5 degrees for nose down trimming.

The wiring schematics, however, just say (a tiny) 1.5 degrees for both directions.

Interestingly, to get this trimming inhibit system to work, the columns have to be moved in the same direction.
You can still trim the aircraft in either direction if the columns are split (at least on later model -300's)

emerydc8

Thanks, John.
Quote

Interestingly, to get this trimming inhibit system to work, the columns have to be moved in the same direction.

Do you think this is a typo?

John H Watson

It's shown in the wiring diagrams.

The trimming signal from the active SAM (Stab Trim/Aileron Lockout Module) goes via both cutout switches (at the bases of the columns) to the tail. The switches are in parallel. You can have one switch in the cutout position (open circuit), but the trimming signal can still go via the other one (closed circuit) if the stick is in neutral or in the opposite direction.

I assume that if the pilots were trimming in opposite directions, the SAM would see this as a fault and not give a command to the stabilizer.

emerydc8

On the classic, after every engine start we would trim the stab aft and push the yoke forward to verify that the stab trim cutout was working. Then do the same in the opposite direction and check it again.

If you have an unscheduled stab trim, you turn both stab cutout switches off, observe the unscheduled stab stopped moving and then turn one switch back on (center) and see if the runaway continues. If it doesn't (I guess you have a 50/50 chance), then you would leave it like it is and use the center stab trim system to bring it back into trim.

So, in the situation below, you would want stab trim movement in the direction of the yoke.

For example, you get an unscheduled stab trim aft and you override with full forward yoke and move both cutout switches off. The stab stops. Then you move the center system cutout switch back on and check for further unscheduled stab movement. If no movement, you would leave the other switch (left) off and bring it back into trim using the center system. So the stab would be moving forward in the direction of the full forward elevator. You wouldn't want the stab trim to move opposite (aft).

Does that make sense?


Hardy Heinlin

QuoteToga didn't engage until after flaps were set to 1

Does the 767 include the same 2-second delay logic that is included in the 744 for the time from TOGA push to THR REF engagement on the ground? If so, maybe the delay is used for some parameter validations, like the flaps-out condition etc. If the system detects any flaps-out status before the timeout is passed, the TOGA command might be executed. Perhaps! I don't know.


|-|ardy


emerydc8

Either that or the captain could have hit the TO/GA switch after selecting flaps 1.